I heard on the radio this morning about a sudden 20-degree temperature rise in Wichita, Kansas (very early on June 9, 2011), and I am wondering how this can happen.
Allison Jones, Bellingham, Washington
I read in this morning's (June 11, 2011) Santa Rosa Press Democrat about strange weather in Wichita, Kansas. The story said the temperature increased from 85°F to 102°F just after midnight due to a pocket of “warm air that had collapsed,” and, when it fell, it created winds up to 50 mph in parts of the city. How can this happen? Warm air is supposed to be lighter than cooler air and rises.
Warren Bilstein, Ukiah, California
Both of you refer to an unusually intense heat burst that occurred about 12:45 a.m. in Wichita Kansas, very early on June 9, 2011. This column first discussed heat bursts in the September/October 1995 issue of Weatherwise, but the 2011 event in Wichita has been so well documented that it is time to revisit this topic. I thank Kenneth Cook, Science and Operations Officer, and Kevin Darmofal, Lead Meteorologist, both at the NWS office in Wichita, Kansas, for documenting this event and checking what I wrote.
First, let's review what probably happens in a heat burst with the aid of Figure 1, repeated from the original 1995 column in Weatherwise. Imagine that we are looking south at the back side of a cluster of thunderstorms that is moving east (from right to left in the figure). The edges of the thundercloud are scalloped. The top of the thundercloud is near 13 km altitude. The anvil, just 120 km of which is shown, extends beyond the right and left edges of the figure. Its lower edge slopes downward from about 6.0 km altitude on the right to about 2.5 km, the altitude of the base of the thundercloud. The base lies just below the freezing level (0°C). The heavy precipitation from this thunderstorm cluster lies off the left edge of the figure, but stratiform (gentle) rain is falling at left, with intensity decreasing from left to right. At the middle of the diagram, rainfall has ceased. Some light precipitation is still falling from the base of the anvil, but it quickly evaporates into the drier air below.
Caption: Figure 1. A conceptual model of the heat burst. A portion of the anvil cloud associated with a cluster of thunderstorms is shown by the scalloped outline in vertical cross-section. The anvil is moving from right to left with the prevailing wind. Air motion within the anvil (bold straight arrow) opposes the prevailing flow. Stratiform precipitation still reaches the ground toward the left, but the heavier thunderstorm showers are off the left edge of the diagram. Mid-tropospheric air, drawn into the storm from the back side, takes a downward-sloping trajectory following the lower edge of the anvil (solid, curved arrow). The trajectory of heat burst air is shown by the dashed arrow. The air associated with the heat burst temporarily scours out a shallow layer of stable air hugging the surface. The dotted lines show the upper boundary of this stable layer. From an article by Bernstein and Johnson, Monthly Weather Review, February 1994.
The airflow in the anvil is shown by the straight arrow pointing to the right at 10 km altitude. Thunderstorms lift great volumes of air all the way to the tropopause (the top of the troposphere). As these updrafts suddenly encounter the stable layer of air at the base of the stratosphere, they lose their buoyancy, that is, they become colder than the air outside the cloud at the same altitude and stop rising. (If the air in the updraft does penetrate the tropopause, it does so only briefly, then sinks back down.) The mass in this great chimney of rising air must go somewhere, and the path of least resistance is for it to spread laterally at the tropopause in the shape of an anvil. The thunderstorm updraft not only blocks the environmental flow, but some of the updraft air actually spreads into the wind, which helps to give the back of the anvil its characteristic shape. That is why the arrow at 10 km in the anvil points west, even though the environmental flow around and over the cloud is in the opposite direction.
Environmental air, entering from the right, is forced downward along the sloping underside of the anvil. Some thunderstorm clusters actually develop low pressure at mid-tropospheric levels; in this case, environmental air may be accelerated toward the back side of the anvil (from the right), creating localized strong winds in mid-troposphere that wouldn't otherwise be there. Light precipitation, falling from the base of the anvil, evaporates into the descending, dry environmental air. If the lapse rate (decrease of temperature with height) is near dry adiabatic1 (1°C per 100 m or 5.4°F per 1000 feet), this rain-cooled air, already descending near the base of the anvil, will become cooler than the subcloud air and accelerate earthward. The descending air warms by compression, 1°C for every 100 m it sinks, and its relative humidity drops.
1The dry adiabatic lapse rate is the rate of cooling that a parcel of unsaturated air (relative humidity less than 100%) experiences when it expands upon being lifted. This lapse rate is often expressed as °C of cooling per 100 m of lift. This cooling rate is realized only if the parcel does not exchange air with its surroundings. Under these same conditions, if an air parcel is forced downward, it will experience an equivalent rate of warming by compression. The dry adiabatic lapse rate is thus defined by any process that lifts unsaturated air or forces it downward. On warm, summer afternoons, the atmosphere may develop a dry adiabatic lapse rate in the lowest kilometer or so when heated from below by warm ground. In such an environment, vertical motion occurs freely because rising and sinking parcels of air always find themselves at the same temperature as their surroundings.
How low can this air go? The receding showers have probably left in their wake a shallow pool of rain-cooled air near the ground, often less than 500 m thick. Even without prior rainfall, since heat bursts are invariably observed at night, there is likely to be a shallow inversion near the ground (warmer air above cooler air). In Figure 1, the dotted line represents the top of this stable layer of air. The down-rushing air (dashed arrow in Figure 1) must have sufficient momentum to penetrate this stable layer of air and scour it out before a heat burst can occur at the surface. Downdrafts of six to eight m s−1 can penetrate shallow stable layers.
Heat bursts are short-lived, usually lasting only a few minutes. In the same way that my puffing very strongly on a shallow puddle temporarily displaces the water, I soon run out of air, and the water flows back. In the same way, a current of down-rushing air, less dense than its surroundings, displaces a shallow layer of stably stratified air. But the downdraft soon loses it punch, and the cooler, stable air flows back.
When the heat burst strikes, the temperature rises rapidly, the dewpoint falls rapidly (and so also the relative humidity), and the wind blows, possibly strongly, driven by the downdraft.
The above scenario is plausible, even probable, for heat bursts, but the exact way in which such strong downdrafts develop is still an open question. Nonetheless, it is evident that heat bursts occur only under very special conditions:
Thunderstorms recently in the vicinity, often in the dissipating stage
Region of stratiform rain toward the rear of the thunderstorm complex
Sloping back side of the anvil extending upwind with respect to the air flow around the cloud
Dry air and steep lapse rate in the air immediately beneath the anvil
Moderate to strong air flow in mid-troposphere along the sloping back side of the anvil
Precipitation evaporating into drier air beneath the anvil
Shallow, stable layer of air near the ground
Theoretically, a heat burst could occur in the daytime, but the surface temperature is usually already high from solar heating before thunderstorms develop. Unless rain-cooled air from a shower significantly lowers the temperature, the subsequent rise in temperature caused by a heat burst won't catch anyone's attention.
Now to the observed heat burst in Wichita, Kansas, very early on the morning of June 9, 2011.
In dissipating thunderstorms, the heavy precipitation has ended, most of the air motion in the cloud is downward, and the cloud is literally collapsing. Figure 2 shows this in four infrared images from the GOES-13 satellite, spaced 30 minutes apart and bracketing the time of the heat burst in Wichita. The image times are 0015, 0045, 0115, and 0145 CDT on June 9. The color code at the top of each frame indicates cloud-top temperature, which is close to -50°C at 0015 CDT but rises to -40°C 90 minutes later. Throughout these 90 minutes, the area of the cold cloud top shrinks. Surface weather data are plotted in Figure 2 for Wichita (KICT_at the center of the image) and other surrounding stations. They are almost impossible to see, but the midnight observation for KICT (84°F) is plotted on image (a), the 0100 CDT observation (101°F) is plotted on images (b) and (c), and the 0200 CDT observation (81°F) is plotted on image (d). These official observations document the heat burst, but they do not indicate how quickly it comes and goes.
Caption: Figure 2. Four infrared images from the GOES-13 satellite spaced at 30-minute intervals and acquired early on June 9, 2011: (a) 0015 CST, (b) 0045 CST, (c) 0115 CST, and (d) 0145 CST. Surface observations are plotted for 0000 CST in (a), for 0100 CST in (b) and (c), and for 0200 CST in (d). Wichita, Kansas (KICT) is at the center of the each image. The areal coverage includes Southeast Kansas. The Kansas-Oklahoma state line runs horizontally across the image near the south end of the thunderstorm anvil.
Fortunately, the Wichita Automated Surface Observing System (ASOS) archives one-minute data, though very few of these data are transmitted over official channels in real time. Figure 3 shows wind direction and speed (with gusts), temperature, dewpoint, and pressure over a six-hour period bracketing the heat burst. A stiff breeze from the south characterized the weather in the few hours before the heat burst. A momentary gust to 60 knots just before 11:00 p.m. may have been a downburst, with a few sprinkles of nearly evaporated rain accompanying a strong downdraft of air from a high-based shower. That, however, was not the main event.
Caption: Figure 3. One-minute surface weather data for Wichita (KICT), Kansas, late on June 8 and early on June 9, 2011. Local time (CST) is indicated. The bold vertical line slicing through the figure indicates the time of the heat burst, characterized by a sudden increase in wind speed and change in direction, a rise in temperature to near-record maximum for the date, a large decrease in dewpoint, and a pressure drop.
At about 12:45 a.m., the wind suddenly shifted from southwest to northeast, gusting to 40 knots. Within a few minutes, the temperature rose from the low 80s to 101°F, just one degree shy of the record maximum for the day, despite the hour – in the middle of the night. The dewpoint plunged from the middle 60s to the upper 20s, suggesting that this air must have descended from considerable altitudes. The sudden drop in pressure and subsequent rise is characteristic of heat bursts. The column of unusully warm air temporarily passing over the station exerts less surface pressure than the air columns surrounding the heat burst. The infrared image in Figure 1b is simultaneous with the heat burst. It is clear that the anvil canopy was still over Wichita at the time of the heat burst, which is consistent with the conceptual model presented earlier.
WSR-88D Doppler radar images at the time of the heat burst (0019 CDT) are shown in Figure 4. Reflectivity is on the left, radial velocity on the right. The Wichita radar is collocated with the NWS office. It lies within the dark circle at the center of each image. The reflectivity display indicates a thunderstorm still in progress, well south of Wichita. Red colors indicate reflectivity values greater than 50 dBZ. Only weak precipitation echoes (blue) remain a few miles southwest of the radar, where the heat burst was first detected. The reflectivity display gives no clue about the heat burst, but the radial velocity display at right does. The latter display indicates the component of wind approaching or receding from the radar. Green and blue colors indicate air approaching the radar; orange and yellow colors indicate air receding from the radar. In this image, there are a few dark blue pixels at the very center of the yellow circle, indicating winds greater than 45 knots approaching the radar. These pixels are almost impossible to see unless one zooms in on the yellow circle, as in Figure 5. The heat burst signature is small, indeed, which gives some idea of the very low probability of experiencing this phenomenon in your own backyard.
Caption: Figure 4. Data from the WSR-88D Doppler radar at Wichita, Kansas, acquired at 0519 UTC (0019 CDT) on June 9, 2011, about the time of earliest detection of the heat burst. The heat burst did not reach the Wichita Airport, where the radar is located, until about 25 minutes later. Left: Reflectivity data. The color bar at top right indicates reflectivity values in dBZ. Right: Radial velocity data. The color bar indicates radial speed in knots. Negative values indicate that air is approaching the radar; positive values indicate that air is receding from the radar. The radar itself is at the center of each image, within the black circle. At the center of the yellow circle (right) are a few blue pixels, indicating heat burst winds approaching 45 knots. These pixels are more easily seen in Figure 5, which is a zoomed image.
Caption: Figure 5. A zoomed view of the heat burst signature in the radial velocity display of the Wichita 88D radar, valid at 0019 CDT, June 9, 2011. At this time, the heat burst covered a very small area, only a kilometer or two wide. The burst is embedded within a larger-scale divergent signature, oriented along a northwest-southeast line. This line (not drawn) separates inbound air (green), closer to the radar, from outbound air (beige and orange), more distant from the radar, at low levels.
Thanks to John Osborn, NOAA Earth System Research Lab, for help in formatting the figures.
Did you know that it's now easier than ever to submit questions to this column? Just use our new email address: email@example.com. If you attach photos, please be sure that they have a resolution of at least 300 dots per inch when they appear on the printed page. Photos taken with 5-megapixel or higher digital cameras will normally satisfy this requirement. Also be sure to include your name and the city or town where you live (or the nearest town if you live in a rural area). Weatherwise cannot answer questions in print without this information.
Weatherwise Contributing Editor THOMAS SCHLATTER is a retired meteorologist and volunteer at NOAA’s Earth System Research Laboratory in Boulder, Colorado. Submit queries to the author at firstname.lastname@example.org, or by mail in care of Weatherwise, Taylor & Francis, 325 Chestnut St., Suite 800, Philadelphia, PA 19106.